2018 ANNUAL DIRECT
OBSERVATION SURVEY OF SAFETY BELT USE
AND MOBILE DEVICE USE
Prepared for: Michigan Office of Highway Safety Planning
Lansing, MI
Prepared by: Michigan State University
East Lansing, MI
Date: July 2018
2018 ANNUAL DIRECT OBSERVATION SURVEY OF SAFETY BELT USE AND MOBILE DEVICE
USE
Prepared for: Michigan Office of Highway Safety Planning
Lansing, MI
Prepared by: Michigan State University
East Lansing, MI
Date: July 18, 2018 The opinions, findings, and conclusions expressed in this publication are those of the author(s) and not necessarily those of the Michigan Office of Highway Safety Planning, the U.S. Department of Transportation, or the National Highway Traffic Safety Administration. This report was prepared in cooperation with the Michigan Office of Highway Safety Planning and the U.S. Department of Transportation, and the National Highway Traffic Safety Administration.
Technical Report Documentation Page
1. Report No.
2. Government Accession No. 3. Recipient’s Catalog No.
4. Title and Subtitle 2018 Annual Direct Observation Survey of Safety Belt and Mobile Device Use
5. Report Date: July 18, 2018
6. Performing Organization Code:
7. Author(s) Timothy J. Gates, Peter T. Savolainen, Brendan J. Russo, Steven Stapleton, and Jonathan J. Kay
8. Performing Organization Report No.
9. Performing Organization Name and Address: Michigan State University 428 S. Shaw Lane Department of Civil and Environmental Engineering East Lansing, MI 48824
10. Work Unit No. (TRAIS)
11. Contract or Grant No.
12. Sponsoring Agency Name and Address: Office of Highway Safety Planning 7150 Harris Drive Dimondale, MI 48821
13. Type of Report and Period Covered: Final Report
14. Sponsoring Agency Code:
15. Supplementary Notes:
16. Abstract: This report documents the results of the 2018 Annual Direct Observation Survey of Safety Belt and Mobile Device Use in the State of Michigan. Safety belt use by drivers and front seat passengers was monitored at a total of 200 intersection/interchange sites within 35 counties throughout Michigan during late May and early June 2018. In addition to belt use, data were collected for vehicle type and use, as well as the gender, age, and race for each observed front seat occupant, and mobile device use for each observed driver. The results of this survey show the weighted safety belt usage rate in the state of Michigan for 2018 is 93.4 percent. This represents a slight decrease from the 94.1 percent use rate observed during the 2017 Annual Direct Observation Survey, though the change is not statistically significant. Males and younger occupants, specifically those in pick-up trucks, continue to exhibit lower belt use rates. The observed rate of hand-held device use by all vehicle drivers is 7.1 percent, which represents an increase from the 6.0 percent device use rate observed during the 2017 Annual Direct Observation Survey.
17. Key Words:
Safety belt use, use rate by vehicle type, mobile device use rate, gender and demographic characteristics
18. Distribution Statement: Unlimited
19. Security Classification (report): Unclassified
20. Security Classification (Page): Unclassified
21. No of Pages:
49
22. Price:
i
TABLE OF CONTENTS PAGE
1.0 INTRODUCTION ..................................................................................................................................... 1
1.1 Study Purpose and Objectives ......................................................................................................... 2 1.2 Study Area ....................................................................................................................................... 2 2.0 SAMPLING METHOD ............................................................................................................................. 2
2.1 General Approach ........................................................................................................................... 2 2.2 Road Segment Stratification ........................................................................................................... 6 2.3 Selection of Road Segments .......................................................................................................... 6 2.4 Selection and Scheduling of Survey Locations ............................................................................. 10 2.5 Data Collection Process ................................................................................................................ 10 2.6 Rescheduling and Alternate Sites ................................................................................................. 12 2.7 Sample Size and Precision ........................................................................................................... 12 3.0 OBSERVER TRAINING ........................................................................................................................ 12 4.0 QUALITY CONTROL ............................................................................................................................ 14 5.0 DATA ANALYSIS .................................................................................................................................. 14
5.1 Imputation ..................................................................................................................................... 15 5.2 Sampling Weights ......................................................................................................................... 15 5.3 Non-Responding Site Adjustment ................................................................................................. 15 5.4 Estimators ..................................................................................................................................... 15 5.5 Variance Estimation ...................................................................................................................... 16 5.6 Non-Response Rate ..................................................................................................................... 17 6.0 RESULTS AND CONCLUSIONS ......................................................................................................... 18
6.1 Safety Belt Survey Results and Conclusions ................................................................................ 18 6.2 Mobile Device Use Results and Conclusions ............................................................................... 29
REFERENCES ............................................................................................................................................ 32
APPENDIX I – Michigan Safety Belt Survey Cover Sheet and Data Collection Form ................................ 33
APPENDIX II – Resumes of Timothy J. Gates and Peter T. Savolainen ................................................... 36
APPENDIX III – List of Annual Observation Locations by County, Stratum, and Road Classification, Including Safety Belt Use Observation Data .................................. 39
ii
LIST OF FIGURES PAGE
Figure 1: 35-County Sample for the Direct Observation Safety Belt Surveys……………………….…….4 Figure 2: Training Syllabus ........................................................................................................................ 13
LIST OF TABLES PAGE
Table 1. Michigan Average Motor Vehicle Crash-Related Fatalities by County (2010-2014) ..................... 5 Table 2. Michigan MAF/TIGER Feature Class Code Codes Included in the Road Segment File ............... 6 Table 3. Roadway Functional Strata by County, Road Segments Population (N), Length of Selected Segments (miles), and Number of Segments Selected (n) ......................................... 8-9 Table 4. Safety Belt Use Codes and Definitions ........................................................................................ 11 Table 5: Annual Vehicle Miles of Travel by Stratum, 2016 (in 1,000’s) ..................................................... 16 Table 6: Annual Weighted Safety Belt Use Rate for Drivers and Front-Seat Passengers ........................ 18 Table 7: Annual Raw/Unweighted Safety Belt Use Summary ................................................................... 19 Table 8: Annual Safety Belt Use Day and Time Sampling Summary ........................................................ 19 Table 9: Annual Safety Belt Use Rates by Stratum and County ................................................................ 20 Table 10: All Vehicles Annual Belt Use Summary ..................................................................................... 21 Table 11: Passenger Cars Annual Belt Use Summary .............................................................................. 22 Table 12: Sport Utility Vehicles Annual Belt Use Summary ....................................................................... 23 Table 13: Vans/Minivan Annual Belt Use Summary .................................................................................. 24 Table 14: Pick-Up Trucks Annual Belt Use Summary ................................................................................ 25 Table 15: Annual Belt Use by Demographic Characteristics ................................................................. 27-28 Table 16: Annual Weighted Mobile Device Use Rate for Drivers ............................................................... 29 Table 17: Annual Unweighted Mobile Device Use Rates by Use Type ...................................................... 29 Table 18: Annual Mobile Device Use Summary .................................................................................... 30-31
1
1.0 INTRODUCTION
The use of safety belts is perhaps the single most effective means of reducing fatal and non-fatal injuries
in motor vehicle crashes. In the first half of 2017 alone, a statistical projection estimated 17,530 people
were killed in motor vehicle crashes in the United States; only a marginal decrease of 0.6 percent compared
with 2016 [1]. Past research indicates that the use of safety belts reduces the risk of fatal injury to front
seat occupants by approximately 45 percent for passenger vehicles and 60 percent for light trucks [2].
Moreover, the use of safety belts reduces the risk of moderate to critical injury by 50 percent for occupants
of passenger vehicles and 65 percent for the occupants of light trucks [2]. In 2016 alone, safety belts saved
approximately 14,668 passenger vehicle occupants over the age of 5 [2]. A recent study conducted by the
National Highway Traffic Safety Administration (NHTSA) on the economic and societal impacts of motor
vehicle crashes states “The comprehensive societal benefits from safety belt use are enormous” [3]. In
fact, this study found that from 1975 to 2010, safety belts have prevented $7.6 trillion in societal harm as
measured by comprehensive costs, and are currently preventing $330 billion in societal harm annually [3].
Therefore, even small increases in safety belt use rates may potentially lead to important societal benefits.
In light of these facts, continuing efforts have been aimed at increasing the use of safety belts across the
United States. According to a 2017 nationwide safety belt survey, 89.7 percent of drivers and right-front
passengers use safety belts, which is a marginal decrease from the 90.1 percent observed in 2016 [4]. The
Midwest region as a whole showed an 88.6 percent safety belt use rate in 2017, an increase from the 85.5
percent safety belt use rate observed in 2016 [4]. In Michigan, past safety belt use studies indicate the
overall use among front seat occupants increased until 2009, prior to a series of gradual declines. Despite
these declines, the 2017 use rate was 93.4 percent, indicating the use rate in Michigan one of 23 states
with safety belt use rates higher than 90 percent [5]. It is important to recognize Michigan is currently one
of the thirty-four “primary law” states, where a front seat occupant motorist can be stopped and cited for the
sole reason of not wearing a safety belt. The most recent available national statistics (2017) indicate that
states with primary safety belt laws exhibited an average use rate of 90.9 percent, which is 5.2 percent
higher than the 85.7 percent exhibited by states without primary safety belt laws [4].
As the non-use of safety belts is ultimately a behavioral issue, targeted programs aimed at changing belt
use behavior of vehicular occupants who are most prone to low belt use rates represent an important tool
towards increasing use rates. To that end, identification of demographic characteristics related to low belt
use is a primary goal of state belt use surveys. Other uses of state safety belt use include:
To fulfill reporting requirements to NHTSA;
To allocate statewide safety funding to specific program areas;
To provide targeted funding to specific areas within the state where use rates are lower than the
statewide average; and
To provide targeted programs for certain segments of the population.
2
1.1 Study Purpose and Objectives
The purpose of this study was to perform the Annual Direct Observation Survey at 200 roadside locations
to determine the percentage of drivers and front-seat passengers who were utilizing their safety belts
correctly and the percentage of drivers using mobile devices. Additional objectives were as follows:
Implement the methodology for estimating Michigan belt use in an economically feasible manner
that is compliant with the Uniform Criteria for State Observational Surveys of Seat Belt Use;
Provide training to all staff conducting the observation surveys and conduct quality
assurance/quality control (QA/QC) of the data collection efforts;
Conduct an observational survey of safety belt use for two weeks in the month of September;
Summarize and cross-tabulate the observational data in a spreadsheet format indicating overall
safety belt use, safety belt use by strata, safety belt use by time of day and day of week, and safety
belt use by various demographic characteristics; and
Continue to track changes in safety belt use and generate necessary comparative data and
analyses to assess the relevancy of the 2018 data and compare results to previous surveys.
1.2 Study Area
The study area for the annual observational survey included those counties representing at least 85 percent
of the passenger vehicle fatalities according to Fatality Analysis Reporting System (FARS) data averages
for the years 2010 to 2014, which was the data analysis period required for site re-sampling in 2017.
Michigan is comprised of 83 counties, 39 of which account for at least 85 percent of the passenger vehicle
crash-related fatalities according to FARS data averages for the years 2010 to 2014. Therefore, observation
locations from within these 39 counties were eligible to be selected for inclusion in the survey. As required
by NHTSA, Michigan will update the sample of data collection sites every five years in order to have survey
results that represent the geographic areas with at least 85 percent of crash-related fatalities.
2.0 SAMPLING METHOD
In 2011, the National Highway Traffic Safety Administration (NHTSA) issued new Uniform Criteria for State
Observational Surveys of Seat Belt Use in Federal Register Vol. 76, No. 63 (April 1, 2011, Rules and
Regulations, pp. 18042 – 18059). The current methodological approach was prepared for the State of
Michigan as a part of the 2013 direct observation safety belt survey and was subsequently approved by
NHTSA. The methodology was employed during the sampling of locations used in the surveys performed
during the five-year period of 2013 through 2017. However, the federal criteria also requires that states re-
sample the observation locations using the approved methodology at least every five years. Thus, the 200
primary and 200 alternative observation sites were re-sampled for the 2018-2022 state of Michigan safety
belt surveys. This re-sampling task was performed by Michigan State University based on the NHTSA-
approved methodology for the state of Michigan (developed in 2013), using updated FARS and vehicle
3
miles traveled (VMT) data. The methodology and lists of 200 primary and 200 alternative sites for the 2018-
2022 surveys were approved by NHTSA in early 2018. Please refer to Appendix II for the resumes of the
principal investigators, Dr. Timothy Gates and Dr. Peter Savolainen, who in addition to leading the re-
sampling effort for the FY2018-2022 surveys, also led development of the methodological approach for the
state of Michigan as a part of the FY 2013 safety belt survey. The following sections provide details of the
sampling process.
2.1 General Approach
The study approach includes a stratified systematic probability proportional to size (PPS) sample of data
collection sites as described here:
1. All 83 counties in Michigan were listed in descending order of the average number of motor vehicle
crash-related fatalities for the period from 2010 to 2014. FARS data were used to determine the
average number of crash-related fatalities per county. It was determined 39 counties accounted for
at least 85 percent of Michigan’s total crash-related fatalities during this period as shown in Table
1. These 39 counties comprise the sample frame.
2. The counties were stratified according to historical safety belt use rates into four strata. These
strata were constructed such that the annual vehicle miles of travel (VMT) were approximately
balanced within each of the four groups. This represents the first stage of sample selection.
3. At the second stage, the MAF/TIGER Feature Class Code (MTFCC, see Section 2.2) was used to
classify all road segments into three explicit classifications: 1.) Primary Roads, 2.) Secondary
Roads, and 3.) Local Roads. This resulted in a total of 12 strata (4 belt use strata, each with 3
MTFCC classes). The number of sites within each MTFCC class was determined proportionately
based upon historical VMT, resulting in 30 percent primary roads, 60 percent secondary roads, and
10 percent local roads.
4. Road segments were then implicitly stratified by county and segment length. Specific segments
were selected randomly with PPS from all segments within each stratum. A random, systematic
sample of 50 road segments was selected PPS to road segment length within each belt use group.
This process resulted in the selection of 200 road segments (4 belt use rate groups x 50 sites per
belt use rate group, allocated proportionately among MTFCC classes). An additional 200 sites
were also selected to use as alternates. Figure 1 shows a map displaying the 35-county sample for
the annual direct observation safety belt survey.
4
5. It was initially expected each site would result in a sample size of approximately 125 vehicles,
resulting in approximately 25,000 vehicle observations overall based upon past experience with
the Michigan Annual Safety Belt Use Study. Based on these figures, the standard error was
expected to be less than 2.5 percent. In the event the calculated standard error should be greater
than 2.5 percent, additional data would be collected from existing sites until this criterion was
satisfied.
6. Additional stages of selection were used to determine travel direction, lane, day of week, time of
day, and vehicles to be observed, at random and with known probability, as appropriate under the
Uniform Criteria, as described in Section 2.4.
Figure 1: 35-County Sample for the Direct Observation Safety Belt Surveys
5
Table 1. Michigan Average Motor Vehicle Crash-Related Fatalities by County (2010-2014)
County Average Annual Fatalities (FARS)
Fatality Percentage Within Michigan
Cumulative Fatality Percentage
WAYNE 158.0 16.9% 16.9%
OAKLAND 60.6 6.5% 23.3%
KENT 50.4 5.4% 28.7%
MACOMB 48.8 5.2% 33.9%
GENESEE 36.2 3.9% 37.8%
WASHTENAW 28.2 3.0% 40.8%
MONROE 26.4 2.8% 43.6%
KALAMAZOO 25.4 2.7% 46.3%
BERRIEN 20.8 2.2% 48.5%
SAGINAW 20.4 2.2% 50.7%
INGHAM 19.4 2.1% 52.8%
ST. CLAIR 18.6 2.0% 54.8%
OTTAWA 18.0 1.9% 56.7%
LIVINGSTON 17.2 1.8% 58.5%
MUSKEGON 16.8 1.8% 60.3%
JACKSON 16.6 1.8% 62.1%
CALHOUN 14.4 1.5% 63.6%
ALLEGAN 14.0 1.5% 65.1%
BAY 13.4 1.4% 66.5%
LENAWEE 13.2 1.4% 67.9%
VAN BUREN 12.8 1.4% 69.3%
GRAND TRAVERSE 11.4 1.2% 70.5%
EATON 10.6 1.1% 71.6%
BARRY 10.2 1.1% 72.7%
MONTCALM 9.8 1.0% 73.8%
LAPEER 9.6 1.0% 74.8%
ST. JOSEPH 9.6 1.0% 75.8%
CASS 9.2 1.0% 76.8%
TUSCOLA 9.2 1.0% 77.8%
IONIA 9.0 1.0% 78.8%
ISABELLA 8.2 0.9% 79.6%
NEWAYGO 7.8 0.8% 80.5%
CLINTON 7.3 0.8% 81.2%
HILLSDALE 7.2 0.8% 82.0%
MIDLAND 7.2 0.8% 82.8%
WEXFORD 7.0 0.7% 83.5%
MECOSTA 6.8 0.7% 84.2%
BRANCH 5.8 0.6% 84.9%
MARQUETTE 5.8 0.6% 85.5%
6
2.2 Road Segment Stratification
Using 2016 Topologically Integrated Geographic Encoding and Referencing (TIGER) data developed by
the U.S. Census Bureau, a comprehensive list of road segments from within these 39 counties was created.
Each of these road segments has been classified by the U.S. Census Bureau using the MAF/TIGER
Feature Class Code (MTFCC). There are primarily three classifications: 1) Primary Roads, 2) Secondary
Roads, and 3) Local Roads (See Table 2 for detailed definitions). In addition, the listings include segment
length as determined by TIGER. This descriptive information allowed for stratification of road segments. A
systematic probability proportional to size (PPS) sample was employed to select the road segments to be
used as observation sites.
Table 2. Michigan MAF/TIGER Feature Class Code Codes Included in the Road Segment File
Code Name Definition
S1100 Primary Road
Primary roads are generally divided, limited-access highways within the interstate highway system or under state management, and are
distinguished by the presence of interchanges. These highways are accessible by ramps and may include some toll highways.
S1200 Secondary
Road
Secondary roads are main arteries, usually in the U.S. Highway, State Highway, or County Highway system. These roads have one or more lanes of traffic in each direction, may or may not be divided, and usually have at-grade intersections with many other roads and driveways. They often have
both a local name and a route number.
S1400
Local Neighborhood Road, Rural Road, City
Street
These are generally paved non-arterial streets, roads, or byways that usually have a single lane of traffic in each direction. Roads in this feature class may be privately or publicly maintained. Scenic park roads would be
included in this feature class, as would (depending on the region of the country) some unpaved roads.
2.3 Selection of Road Segments
Within each of the four belt use strata, a total of 50 road segments were selected. Michigan employed the
Census TIGER EDGES data set for the selection of road segments. Michigan exercised the available
exclusion option and removed rural local roads in counties not within metropolitan statistical areas (MSAs),
and other non-public roads, unnamed roads, unpaved roads, vehicular trails, access ramps, cul-de-sacs,
traffic circles, and service drives from the dataset. The number of road segments selected within each
MTFCC class was determined proportionately based upon total annual VMT within the three classes
(Primary, Secondary, and Local). Thus, the segments selected ultimately included 15 primary roads (20
percent of sample), 30 secondary roads (60 percent of sample), and 5 local roads (10 percent of sample).
Prior to selecting the specific observation locations, all road segments were explicitly stratified by MTFCC
(primary, secondary and local) within each of the four belt use rate groups and implicitly stratified by county
and by segment length to obtain an ordered list. Implicit stratification by county was done to ensure
adequate geographic coverage was obtained as a part of the selection process. Similarly, the implicit
7
stratification by length ensured representative coverage within each MTFCC class since higher-class roads
tended to be longer than lower-class roads. Specific road segments were then selected with PPS using
segment length as the measure of selection (MOS). As such, the inclusion probability for a specific road
segment is:
| / ∑∀ ,
where is the road segment sample size for MTFCC c in stratum that was allocated, is the length
of road segment h, and
∀
is the total length of all segments in stratum and MTCFF c.
A random start (RS) was selected between 0 and the calculated I, which determined the first road segment
selected. Subsequent road segments selected were determined by adding multiples of I to the RS until the
desired number of road segments were selected and/or the end of the sorted list was reached.
Table 3 presents summary statistics detailing the number of eligible road segments (N), the total length
(miles) of these segments, and the number of road segments selected (n) within each of the MTFCC classes
by belt use group and county. Appendix III presents the complete list of the final observation sites including
belt use stratum, county, and road classification.
In the event an original road segment was permanently unavailable, a reserve road segment was to be
used. The reserve road segment sample consisted of one additional road segment per original road
segment selected, resulting in a reserve sample of an additional 200 road segments. These reserve
segments were identified and selected as the road segments immediately following the original road
segment actually selected. Thus, these segments were also explicitly stratified by safety belt use and
MTFCC group, as well as implicitly stratified by segment length and county. Each reserve segment
corresponded to an original road segment actually selected. Thus, these are considered selected with PPS
using road segment length as MOS by the same approach as described previously. As such, for the
purposes of data weighting, the reserve road segment inherited all probabilities of selection and weighting
components up to and including the road segment stage of selection from the original road segment actually
selected. Probabilities and weights for any subsequent stages of selection (e.g., the sampling of vehicles)
would be determined by the reserve road segment itself.
8
Table 3. Roadway Functional Strata by County, Road Segments Population (N), Length of Selected Segments (miles), and Number of Segments Selected (n)
Stratum County Type MTFCC Classification
Primary Secondary Local Total
Stratum 1
Ingham N 272 1203 15017 16492
Length 63 158 1967 2189 n 3 8 0 11
Kalamazoo N 160 729 14749 15638
Length 50 123 2023 2196 n 2 5 0 7
Oakland N 792 1907 65290 67989
Length 164 234 6804 7203 n 7 8 3 18
Washtenaw N 282 910 18992 20184
Length 66 162 2614 2842 n 3 9 2 14
Stratum 2
Allegan N 170 614 11226 12010
Length 58 131 2249 2438 n 2 4 0 6
Bay N 200 726 8954 9880
Length 57 120 1363 1539 n 1 2 0 3
Calhoun N 388 775 10407 11570
Length 120 104 1848 2072 n 4 0 1 5
Eaton N 255 714 7584 8553
Length 78 129 1457 1664 n 3 3 0 6
Grand Traverse N 0 604 8996 9600
Length 0 105 1325 1430 n 0 2 0 2
Jackson N 215 827 11597 12639
Length 61 154 1942 2157 n 3 6 1 10
Kent N 438 1524 33635 35597
Length 88 266 3911 4265 n 0 7 1 8
Livingston N 239 523 14418 15180
Length 61 104 2043 2209 n 1 1 1 3
Midland N 0 461 7172 7633
Length 0 97 1282 1379 n 0 2 0 2
Monroe N 324 740 10324 11388
Length 68 133 1676 1877 n 0 2 1 3
Ottawa N 205 819 15925 16949
Length 70 135 2239 2445 n 1 1 0 2
Stratum 3
Berrien N 447 1059 15481 16987
Length 103 168 2051 2321 n 2 1 0 3
Branch N 108 287 5159 5554
Length 45 52 1219 1316 n 1 0 0 1
Cass N 0 649 5870 6519
Length 0 127 1186 1313 n 0 2 1 3
9
Table 3 - Roadway Functional Strata by County, Road Segments Population (N), Length of Selected Segments (miles), and Number of Segments Selected (n) (Continued)
Stratum County Type MTFCC Classification
Primary Secondary Local Total
Stratum 3
Clinton N 188 369 6505 7062
Length 56 98 1387 1540 n 0 2 0 2
Genesee N 664 802 24988 26454
Length 139 136 2918 3193 n 2 3 0 5
Hillsdale N 0 488 5533 6021
Length 0 113 1365 1478 n 0 1 0 1
Ionia N 164 391 6229 6784
Length 51 78 1334 1463 n 0 2 0 2
Lapeer N 159 382 7611 8152
Length 49 80 1618 1747 n 1 1 1 3
Lenawee N 0 878 2672 3550
Length 0 162 264 425 n 0 2 0 2
Marquette N 0 897 8662 9559
Length 0 184 1639 1822 n 0 3 0 3
Mecosta N 0 446 6597 7043
Length 0 108 1398 1506 n 0 1 0 1
Montcalm N 0 616 8736 9352
Length 0 132 1842 1975 n 0 2 2 4
Saginaw N 307 1047 15814 17168
Length 61 170 2390 2621 n 3 1 0 4
St. Clair N 388 865 11924 13177
Length 107 107 1987 2201 n 2 0 0 2
St. Joseph N 0 831 6885 7716
Length 0 140 1277 1417 n 0 1 1 2
Tuscola N 0 651 408 1059
Length 0 141 39 180 n 0 2 0 2
Van Buren N 198 450 8193 8841
Length 75 85 1618 1777 n 4 4 0 8
Wexford N 0 680 5235 5915
Length 0 155 1119 1274 n 0 2 0 2
Stratum 4
Macomb N 402 1651 39648 41701
Length 65 159 3745 3970 n 3 14 3 20
Wayne N 2041 3860 85981 91882
Length 250 292 7620 8161 n 12 16 2 30
10
2.4 Selection and Scheduling of Survey Locations
Road segments were mapped according to the latitude and longitude of their midpoints. The selected road
segment was identified by an intersection or interchange that occurred within or just beyond the segment.
Data collection sites were deterministically selected such that traffic would be moving during the observation
period. Therefore, sites were assigned to locations within the segment that were 50 to 150 feet from any
controlled intersections. For limited access roadways, data collection occurred on a ramp carrying traffic
exiting the highway. The observed direction of travel was randomly assigned for each road segment.
All belt use observations were conducted during weekdays and weekends between 7 AM and 7 PM to
include rush hour (before 9:30 AM and after 3:30 PM) and non-rush hour observations. Site assignment
schedules, which were provided to the data collectors and quality control monitors, indicated the observed
road name, nearest crossroad, GPS coordinates where the observer should stand, assigned date, assigned
time, and assigned observation direction. Sites within relatively close geographic proximity were assigned
as data collection clusters. In accordance with the uniform safety belt survey criteria, the first site within
each cluster was assigned a random day and time for completion. All other sites within a cluster were
assigned to the same day and by geographic proximity to minimize travel within the cluster. Approximately
five sites were scheduled each day for each data collector. Start times and days were staggered to ensure
all days of the week and hours of the day (daylight) were represented in the sample.
2.5 Data Collection Process
Safety belt surveys were performed for exactly 60 minutes at each of the 200 observation locations. Wayne
State University (WSU), under subcontract to MSU, collected data at those study sites in Wayne, Oakland,
Macomb, and Monroe Counties, while MSU collected data at all other locations. The data collected at the
200 observation sites provided a representative sample for each day of the week and each hour of the day
between 7 AM and 7 PM of the statewide safety belt use characteristics. All passenger vehicles, including
commercial vehicles weighing less than 10,000 pounds, were eligible for observation. Heavy truck, buses,
and other vehicles weighing over 10,000 pounds were not observed. Only one direction of traffic was
observed at any given site. The data collectors were instructed to observe as many lanes of traffic as they
could while obtaining data on 99 percent of eligible vehicles. This direction of observation was pre-
determined at each location as explained previously. The observations were appropriately weighted, as
explained in the Data Analysis Section of this report (Section 5.0).
The observers carried a cover sheet and numerous safety belt observation data collection paper forms to
each site. These forms are shown in Appendix I. The observation form was used to record safety belt use
by drivers and front seat passengers, including children in booster seats. The only front seat occupants
excluded from this study were children seated in child seats with harness straps. Table 4 lists the three
clearly defined categories of safety belt use that were observed by the data collectors, which included
11
‘belted correctly’, ‘not belted correctly’, and ‘unknown belt use’ as previously described. An occupant was
recorded as ‘belted correctly’ only if they were observed to be properly using the shoulder belt (i.e. shoulder
belt was across chest; not under arm or behind back). The ‘unknown belt use’ category was marked if an
observer was unable to determine the position of an occupant’s safety belt, and these observations were
not included in the final sample but a record was kept to calculate the non-response rate which is discussed
in the data analysis section of this report.
Table 4. Safety Belt Use Codes and Definitions
Code Definition
Belted The shoulder belt is in front of the person's shoulder and used correctly.
Not belted
The shoulder belt is not in front of the person's shoulder or not used at all.
Unknown It cannot reasonably be determined whether the driver or right front
passenger is belted.
Additional data collected for each observed front-seat occupant included occupant age (estimated), gender,
and race, as well as vehicle type and use (e.g. commercial or non-commercial) information. The driver age
categories included 16-29, 30-59, 60 and over, and unknown, while the passenger age also included a 0-
15 category. The driver and passenger race categories included white, black, other, or unknown. Each
observed vehicle was categorized into one of four groups: passenger cars, sport utility vehicles, vans or
minivans, and pick-up trucks. The vehicles were also identified as commercial or non-commercial vehicles.
Furthermore, the driver was also observed for any indication of mobile device use. The categories included
‘hand-held (talking)’, ‘hand-held (typing)’, ‘hands-free (ear piece)’, and hands-free (no ear piece)’.
The cover sheet was used to document site information, including: date, site location, site number, alternate
site data, assigned traffic flow, number of lanes available and observed, start and end times for
observations, and weather conditions. This cover sheet was completed by the data collector at each site
before any observations took place.
Observations were manually recorded in the field on survey forms and returned back to the office within 24
hours of the data collection, or as soon as possible after multiple day trips to outstate locations. The data
collected in the field were entered into a spreadsheet by the observer at the conclusion of the data collection
activities for each day and verified for accuracy in the office by office staff.
Data collectors also used a hand-held tally device to simultaneously count every passenger vehicle that
passed through the observed lanes during the 60-minute observation period, regardless of whether a safety
belt observation was performed. This volume count was then utilized during the belt use weighting process.
12
2.6 Rescheduling and Alternate Sites
If a site was temporarily unavailable due to a crash, short-term road work or maintenance, inclement
weather, or any event that may hinder exact results, data collection was rescheduled for a similar time of
day and type of day of the week. In the event the site was permanently unavailable, such as being located
within a gated community or closed for long-term construction, then an alternate site selected as part of the
reserve sample was to be used as a permanent replacement.
2.7 Sample Size and Precision
A standard error of less than 2.5 percent for the safety belt use estimates is required by the Final Rule.
Since 1999, Michigan has conducted the Michigan Annual Safety Belt Use Study, and has historically
obtained standard errors below this threshold (e.g. most recently 0.2 percent in 2017) via observed sample
sizes of approximately 25,000 vehicles. Since the proposed design for the 2018 Annual survey was similar
to the 2017 survey, it was expected that the sample size for the 2018 Annual Survey would be similar to
the 2017 Annual Survey and the precision objective was expected to be achieved. In the event that the
precision objective was not met, additional observations would be taken starting with those sites having the
fewest observations. New data would be added to existing data until the desired precision was achieved.
3.0 OBSERVER TRAINING
The data collection team was comprised of MSU and WSU student staff, many of whom have participated
in prior safety restraint use surveys. All data collectors were able to stand for long periods of time, work
outdoors, and successfully complete the training program. The data collector training program included
both a classroom and field portion. The classroom training program was conducted at MSU approximately
three weeks prior to the start of the survey and was led by the PI, Timothy Gates. All data collectors from
both MSU and WSU attended this classroom session. Each data collector received a training manual
composed of the information detailed during the training session and all necessary field supplies. The
syllabus for the training program is shown as Figure 2.
At the conclusion of the classroom training, the data collectors conducted their first field practice at a
location near the MSU campus. QC monitors were available during this period to respond to questions and
offer assistance to data collectors as needed. Reliability and repeatability field data collection practice
continued during the weeks leading up to full-scale survey implementation at various intersections near the
MSU and WSU campuses. These intersections represented various site characteristics that could be
challenging for observational data collection. Initially, inexperienced observers were paired with
experienced observers, who noted which individual vehicle the entire group was to evaluate. This allowed
an analysis of the accuracy of the inexperienced data collectors in comparison to those who have
participated in the study previously. After gaining ample experience, observers were then randomly divided
13
into groups and assigned to collect safety belt observational data independently. The training data was
then entered and compared among the observers in each group to determine the accuracy of their
observations.
Figure 2. Training Syllabus
14
4.0 QUALITY CONTROL
The policies and procedures utilized while conducting the direct observation surveys of safety belt use were
based upon the Uniform Criteria for State Observational Surveys of Seat Belt Use from Title 23, Part
1240.12 of the Code of Federal Regulations. The study design for the Annual Survey was consistent with
these criteria, which established observations should be conducted on specific dates and times and in
particular directions of travel, all of which were determined randomly in advance of the studies. Further,
the criteria state policies should be in place in the event observations cannot be made due to unanticipated
events, such as road construction. In such situations, data collectors were instructed to observe at the pre-
assigned alternate location. Policies were also established for cases where traffic flow is too heavy to
observe all vehicles or traffic is moving too quickly for observation. In most instances, high traffic volumes
prohibit data collectors from observing all vehicles. Consequently, data collectors were instructed to
observe as many vehicles as is feasible for observation under such conditions for the required time period
of 60 minutes, although all passenger vehicles traveling through the observed lanes during the data
collection period were included in the volume count.
The principal investigators from MSU and WSU served as the QC monitors, conducting site audits of the
data collectors. The QC monitor made unannounced covert visits to five percent of all data collection sites
over the duration of the study, which amounted to 10 sites. The purpose of these visits was to ensure data
collectors were following all survey protocol including: performing observational surveys at the assigned
location, in the assigned direction, during the assigned time period, completing the cover sheet and
observation forms correctly, making accurate observations of safety belt use within an appropriate number
of lanes. The random checks were conducted at least once for each observer and no major violations of
policies or procedure were observed as a part of these audits. The QC monitors also checked a 10 percent
random sample of the entered data to ensure the observation data were being entered correctly from the
data collection forms. After data entry, all forms were organized, boxed, and stored for 3-years.
5.0 DATA ANALYSIS
The data collected in the field as a part of the 35-county annual survey were entered into a spreadsheet by
the observer at the conclusion of the data collection activities for each day and verified for accuracy by
office staff. Rates for safety belt and mobile device use were determined for each survey stratum, county,
location, etc., as well as the statewide annual average. A 95-percent confidence interval for each use rate
estimate was determined according to the NHTSA guidelines. The following sections outline the methods
used to estimate the use rate and variance for safety belts. A similar procedure was utilized to estimate
mobile device use rate and variance.
15
5.1 Imputation
No imputation was done on missing data.
5.2 Sampling Weights
The following is a summary of the notation used in this section.
g – Subscript for belt use group strata h – Subscript for road segment strata i – Subscript for road segment j – Subscript for time segment k – Subscript for road direction l – Subscript for lane m – Subscript for vehicle n – Subscript for front-seat occupant
Under this stratified multistage sample design, the inclusion probability for each observed vehicle was the
product of selection probabilities at all stages: for belt use group (stratum-road class), | for road
segment, | for time segment, | for direction, | for lane, and | for vehicle. So
the overall vehicle inclusion probability was:
| | | | | .
The sampling weight (design weight) for vehicle m is: 1
5.3 Non-Responding Site Adjustment
There were no sites which required ‘non-responding’ adjustment in the 2018 Annual Direct Observation
Survey of Safety Belt Use. It should be noted that no observations were recorded at site 152 (Riker Road
and Island Lake Road in Washtenaw County), however since there were no ‘vehicle not observable’ or
‘unknown belt use’ observations here, no non-responding adjustment is required as per An Example of a
Compliant State Seat Belt Use Survey Design [6].
5.4 Estimators
Noting all front-seat occupants were observed, the driver/passenger safety belt use status was:
1, 0,
16
In order to most accurately estimate the weighted safety belt use rate for the entire state of Michigan, the
estimator used in this analysis was weighted by segment length and stratum-level VMT to determine the
overall annual belt use rate in Michigan. This estimation technique is detailed in An Example of a Compliant
State Seat Belt Use Survey Design [6]. Under this estimator, the use rates within each stratum were first
calculated using the road segment length based estimator:
∑ |
∑ |
The twelve stratum-specific use rates were then weighted by the proportion of total statewide VMT (shown
in Table 5) within each stratum, which resulted in the road class VMT-based estimator (pVMT):
∑ ∑∑ ∑
Table 5. Annual Vehicle Miles of Travel by Stratum, 2016 (in 1,000s)
Belt Use Stratum
Road Class Total Primary Secondary Local
1 7,806,660 11,446,321 2,222,223 21,475,204
2 8,142,476 12,010,771 1,810,510 21,963,757
3 5,622,677 11,600,045 1,926,513 19,149,235
4 7,904,825 11,749,975 2,369,860 22,024,660
Statewide 29,476,638 46,807,112 8,329,106 84,612,856
The use of the VMT-based estimator (pVMT) reduced the weighting bias towards local road observation sites
by accounting for their relatively short length and low VMT as compared to primary and secondary roads.
VMT data were obtained from the Michigan Highway Performance Monitoring System (HPMS) for the year
2016.
5.5 Variance Estimation
The variance (and standard error) for each estimator was determined using the “Delete-1 Jackknife”
variance estimation program in SUDAAN 11 software. Under this methodology, the variance was
calculated by deleting one observation location and adjusting the weights of the remaining PSU’s in the
same stratum to account for the deleted PSU. The procedure was repeated, removing each location once.
For the road class VMT based estimator (pVMT), the “Delete-1 Jackknife” method was used to estimate the
variances within each of the road class/belt use strata:
17
1 ′
where:
V p = Estimated variance within each of the road class/belt use strata
= Estimated belt use rate
= Estimated belt use rate at location i in road segment type h in belt use group g
= Estimated belt use rate in road segment type h in belt use group g
= Number of locations of road segment type h in belt use group g
The variance for the annual use rate was then determined using the following equation:
∑∀ ,∀
∑∀ ,∀
where:
= Estimated variance of statewide belt use rate
The standard error of the statewide use rate was found by simply taking the square root of the estimated
variance. The 95 percent confidence interval of the statewide belt use was equal to the weighted safety
belt use rate plus/minus 1.96 (for the Z-test at alpha = 0.05) multiplied by the standard error expressed as
a percent.
5.6 Non-Response Rate
According to NHTSA’s guidelines, the non-response rate for the annual safety belt survey cannot exceed
10 percent. A non-response occurs when the observer was not able to determine the safety belt use of a
front seat vehicle occupant. This can occur due to a variety of reasons such as tinted windows, sun glare,
high speeds of the vehicle in question, etc. Observers in the field marked either ‘vehicle not observable’ or
‘unknown belt use’ to keep a record of the non-response rate. There were a total of 298 non-response
observations which represents approximately 1.4 percent of the total number of observations. This non-
response rate was below the allowable maximum of 10 percent established by the NHTSA.
18
6.0 RESULTS AND CONCLUSIONS
The Annual Direct Observation Survey was performed between Tuesday, May 29, and Tuesday, June 12,
2018. During this observation period, a total of 16,753 vehicles were observed resulting in 20,552 driver
and right-front passenger observations at the 200 observation sites randomly selected to represent
statewide safety belt use according to the federal Uniform Criteria.
6.1 Safety Belt Survey Results and Conclusions
The overall weighted annual safety belt use rate for Michigan in 2018 was found to be 93.4 percent and is
shown in Table 6. The overall weighted annual safety belt use rate was calculated based upon the
procedure described in the Data Analysis section (Section 5.0) of this report. When the safety belt usage
rates were calculated, belted occupants included all drivers and front-seat passengers who were belted
correctly. The “not belted” occupants included drivers and front-seat passengers who were not belted or
who were wearing the belt incorrectly; either under their arm or behind their back. Details of the
observations on an intersection level are provided in Appendix III. It should be noted that there were two
instances where the original site was unobservable (sites 139 and 141), and the assigned alternate sites
were used for these locations as per the study design.
Table 6. Annual Weighted Safety Belt Use Rate for Drivers and Front-Seat Passengers
Observational Wave Safety Belt Use Rate* Standard Error
Annual 93.4% 0.8% 0.4%
* Weighted Safety Belt Usage 95% Confidence Band
The overall annual use rate displayed in Table 6 is representative of all front seat occupants (drivers and
right-front passengers), all daytime hours (7:00 AM-7:00 PM) and all days of the week. Table 7 shows the
raw (unweighted) safety belt use information separated by drivers and front-right passengers. Table 8
summarizes the descriptive statistics for the safety belt survey in terms of sampling statistics for day of the
week and time of the day.
19
Table 7. Annual Raw/Unweighted Safety Belt Use Summary
Belt Use Actual Total No. of Observations
Actual Belted No. of Observations
% Safety Belt Use
Drivers 16,738 15,628 93.4% Passengers 3,814 3,582 93.9%
Total 20,552 19,210 93.5%
Table 8. Annual Safety Belt Use Day and Time Sampling Summary
Day of the Week
Annual Safety Belt Observations
No. of Sites Observed
Percent of Sites in Day of Week
Actual Total No. of Observations
(Occupants)
Percent of Observations in
Day of Week (Occupants)
Sunday 23 11.5% 2661 12.9% Monday 46 23.0% 4153 20.2% Tuesday 33 16.5% 3477 16.9%
Wednesday 25 12.5% 2619 12.7% Thursday 25 12.5% 2648 12.9%
Friday 18 9.0% 2094 10.2% Saturday 30 15.0% 2900 14.1%
Total 200 100.0% 20,552 100.0%
Time of the Day
Annual Safety Belt Observations
No. of Sites Observed
Percent of Sites in Time of Day
Actual Total No. of Observations
(Occupants)
Percent of Observations in
Day of Week (Occupants)
7 am – 8 am 10 5.0% 711 3.5% 8 am – 9 am 14 7.0% 1,153 5.6%
9 am – 10 am 13 6.5% 1,104 5.4% 10 am – 11 am 18 9.0% 1,785 8.7% 11 am – 12 pm 21 10.5% 2,266 11.0% 12 pm – 1 pm 24 12.0% 2,660 12.9% 1 pm – 2 pm 29 14.5% 2,551 12.4% 2 pm – 3 pm 21 10.5% 1,925 9.4% 3 pm – 4 pm 16 8.0% 2,093 10.2% 4 pm – 5 pm 14 7.0% 1,598 7.8% 5 pm – 6 pm 11 5.5% 1,429 7.0% 6 pm – 7 pm 9 4.5% 1,277 6.2%
Total 200 100.0% 20,552 100.0%
The safety belt use rate can be described by the overall use rate, as well as by vehicle type and various
demographics. It should be noted the overall safety belt use rates presented in Table 7 and Tables 9
through 15 represent the raw (un-weighted) safety belt use data. These rates vary from the weighted annual
use rate presented in Table 6. Table 9 summarizes the annual driver and front-seat passenger safety belt
use rates by county and belt-use stratum. Because of the relatively low number of sites and/or observations
in many counties, the safety belt use rates listed may not be fully representative of each county.
20
Table 9. Annual Safety Belt Use Rates by Stratum and County
STRATUM 1 Actual Total No. of Observations
Actual Belted No. of Observations
% Safety Belt Use
Ingham County 1,125 1,045 92.9% Kalamazoo County 900 846 94.0%
Oakland County 2,318 2,144 92.5% Washtenaw County 1,132 1,037 91.6%
Total 5,475 5,072 92.6%
STRATUM 2 Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
Allegan County 576 554 96.2% Bay County 288 270 93.8%
Calhoun County 262 248 94.7% Eaton County 596 542 90.9%
Grand Traverse County 385 369 95.8% Jackson County 1,149 1,091 95.0%
Kent County 1033 972 94.1% Livingston County 322 316 98.1% Midland County 214 197 92.1% Monroe County 352 330 93.8% Ottawa County 188 178 94.7%
Total 5,365 5,067 94.4%
STRATUM 3 Actual Total No. of
Observations. Actual Belted No. of
Observations % Safety Belt Use
Berrien County 264 243 92.0% Branch County 78 76 97.4% Cass County 270 252 93.3%
Clinton County 119 110 92.4% Genesee County 518 469 90.5% Hillsdale County 96 86 89.6%
Ionia County 269 249 92.6% Lapeer County 232 208 89.7%
Lenawee County 267 244 91.4% Marquette County 599 578 96.5% Mecosta County 43 42 97.7%
Montcalm County 151 146 96.7% Saginaw County 450 426 94.7% St. Clair County 271 260 95.9%
St. Joseph County 152 142 93.4% Tuscola County 196 183 93.4%
Van Buren County 783 737 94.1% Wexford County 215 207 96.3%
Total 4,973 4,658 93.7%
STRATUM 4 Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
Macomb County 4,973 4,658 93.7% Wayne County 2,092 1,942 92.8%
Total 2,647 2,471 93.4%Grand Total (Unweighted) 20,552 19,210 93.5%
Stratum 2 displayed the highest safety belt use rate at 94.4 percent, followed by Strata 3 and 4. Stratum 1
displayed the lowest safety belt use rate at 92.6 percent. Tables 10 through 14 summarize occupant safety
belt use for drivers and front-seat passengers by vehicle type for each day of the week, time of the day,
gender, age, and race for the Annual Observation Survey.
21
Table 10. All Vehicles Annual Belt Use Summary
Day of the Week All Vehicle Safety Belt Use
Actual Total No. of Observations
Actual Belted No. of Observations
% Safety Belt Use
Sunday 2,661 2,537 95.3% Monday 4,153 3,846 92.6% Tuesday 3,477 3,251 93.5%
Wednesday 2,619 2,393 91.4% Thursday 2,648 2,486 93.9%
Friday 2,094 1,962 93.7% Saturday 2,900 2,735 94.3%
Total 20,552 19,210 93.5%
Time of the Day Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
7 am – 8 am 711 661 93.0% 8 am – 9 am 1,153 1,054 91.4%
9 am – 10 am 1,104 1,042 94.4% 10 am – 11 am 1,785 1,679 94.1% 11 am – 12 pm 2,266 2,113 93.2% 12 pm – 1 pm 2,660 2,449 92.1% 1 pm – 2 pm 2,551 2,410 94.5% 2 pm – 3 pm 1,925 1,797 93.4% 3 pm – 4 pm 2,093 1,958 93.5% 4 pm – 5 pm 1,598 1,508 94.4% 5 pm – 6 pm 1,429 1,335 93.4% 6 pm – 7 pm 1,277 1,204 94.3%
Total 20,552 19,210 93.5%
Vehicle Type Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
Passenger Cars 7,523 7,029 93.4% Sport Utility Vehicles 7,502 7,119 94.9%
Vans/Minivans 1,920 1,820 94.8% Pick-Up Trucks 3,607 3,242 89.9%
Total 20,552 19,210 93.5%
Gender Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
Male 11,173 10,312 92.3% Female 9,331 8,854 94.9%
Unknown 48 44 91.7% Total 20,552 19,210 93.5%
Age Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
0 - 15 335 314 93.7% 16 - 29 5,562 5,112 91.9% 30 - 59 11,537 10,841 94.0%
60+ 3,106 2,931 94.4% Unknown 12 12 100.0%
Total 20,552 19,210 93.5%
Race Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
White 17,230 16,154 93.8% Black 2,226 2,026 91.0% Other 1,060 999 94.2%
Unknown 36 31 86.1% Total 20,552 19,210 93.5%
22
Table 11. Passenger Cars Annual Belt Use Summary
Passenger Cars Safety Belt Use
Day of the Week Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
Sunday 942 897 95.2% Monday 1,566 1,453 92.8% Tuesday 1,334 1,252 93.9%
Wednesday 916 834 91.0% Thursday 906 858 94.7%
Friday 789 737 93.4% Saturday 1,070 998 93.3%
Total 7,523 7,029 93.4%
Time of the Day Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
7 am – 8 am 277 261 94.2% 8 am – 9 am 428 395 92.3% 9 am – 10 am 388 362 93.3%
10 am – 11 am 578 538 93.1% 11 am – 12 pm 835 785 94.0% 12 pm – 1 pm 1,039 963 92.7% 1 pm – 2 pm 906 857 94.6% 2 pm – 3 pm 745 699 93.8% 3 pm – 4 pm 829 766 92.4% 4 pm – 5 pm 554 520 93.9% 5 pm – 6 pm 510 475 93.1% 6 pm – 7 pm 434 408 94.0%
Total 7,523 7,029 93.4%
Gender Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
Male 3,933 3,656 93.0% Female 3,569 3,355 94.0%
Unknown 21 18 85.7% Total 7,523 7,029 93.4%
Age Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
0 - 15 100 96 96.0% 16 - 29 2,692 2,471 91.8% 30 – 59 3,609 3,399 94.2%
60+ 1,113 1,054 94.7% Unknown 9 9 100.0%
Total 7,523 7,029 93.4%
Race Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
White 5,960 5,603 94.0% Black 1,153 1,042 90.4% Other 398 372 93.5%
Unknown 12 12 100.0% Total 7,523 7,029 93.4%
23
Table 12. Sport Utility Vehicles Annual Belt Use Summary
Sport Utility Vehicles Safety Belt Use
Day of the Week Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
Sunday 1,000 964 96.4% Monday 1,410 1,330 94.3% Tuesday 1,231 1,156 93.9%
Wednesday 1,013 945 93.3% Thursday 1,029 982 95.4%
Friday 737 700 95.0% Saturday 1,082 1,042 96.3%
Total 7,502 7,119 94.9%
Time of the Day Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
7 am – 8 am 268 255 95.1% 8 am – 9 am 423 388 91.7% 9 am – 10 am 384 369 96.1%
10 am – 11 am 677 654 96.6% 11 am – 12 pm 827 783 94.7% 12 pm – 1 pm 928 860 92.7% 1 pm – 2 pm 888 849 95.6% 2 pm – 3 pm 663 622 93.8% 3 pm – 4 pm 720 683 94.9% 4 pm – 5 pm 620 596 96.1% 5 pm – 6 pm 596 566 95.0% 6 pm – 7 pm 508 494 97.2%
Total 7,502 7,119 94.9%
Gender Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
Male 3,324 3,110 93.6% Female 4,165 3,997 96.0%
Unknown 13 12 92.3% Total 7,502 7,119 94.9%
Age Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
0 - 15 143 134 93.7% 16 - 29 1,771 1,650 93.2% 30 – 59 4,414 4,209 95.4%
60+ 1,171 1,123 95.9% Unknown 3 3 100.0%
Total 7,502 7,119 94.9%
Race Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
White 6,334 6,031 95.2% Black 732 672 91.8% Other 422 404 95.7%
Unknown 14 12 85.7% Total 7,502 7,119 94.9%
24
Table 13. Van/Minivan Annual Belt Use Summary
Van/Minivans Safety Belt Use
Day of the Week Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
Sunday 266 258 97.0% Monday 422 396 93.8% Tuesday 327 310 94.8%
Wednesday 259 240 92.7% Thursday 218 202 92.7%
Friday 184 177 96.2% Saturday 244 237 97.1%
Total 1,920 1,820 94.8%
Time of the Day Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
7 am – 8 am 47 42 89.4% 8 am – 9 am 113 106 93.8% 9 am – 10 am 126 120 95.2%
10 am – 11 am 191 182 95.3% 11 am – 12 pm 232 216 93.1% 12 pm – 1 pm 258 249 96.5% 1 pm – 2 pm 242 232 95.9% 2 pm – 3 pm 192 184 95.8% 3 pm – 4 pm 196 183 93.4% 4 pm – 5 pm 129 121 93.8% 5 pm – 6 pm 103 100 97.1% 6 pm – 7 pm 91 85 93.4%
Total 1,920 1,820 94.8%
Gender Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
Male 1,068 1,004 94.0% Female 847 811 95.7%
Unknown 5 5 100.0% Total 1,920 1,820 94.8%
Age Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
0 - 15 39 34 87.2% 16 - 29 374 355 94.9% 30 – 59 1,175 1,117 95.1%
60+ 332 314 94.6% Unknown 0 0 N/A
Total 1,920 1,820 94.8%
Race Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
White 1,621 1,539 94.9% Black 177 167 94.4% Other 118 111 94.1%
Unknown 4 3 75.0% Total 1,920 1,820 94.8%
25
Table 14. Pick-Up Trucks Annual Belt Use Summary
Pick-up Truck Safety Belt Use
Day of the Week Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
Sunday 453 418 92.3% Monday 755 667 88.3% Tuesday 585 533 91.1%
Wednesday 431 374 86.8% Thursday 495 444 89.7%
Friday 384 348 90.6% Saturday 504 458 90.9%
Total 3,607 3,242 89.9%
Time of the Day Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
7 am – 8 am 119 103 86.6% 8 am – 9 am 189 165 87.3%
9 am – 10 am 206 191 92.7% 10 am – 11 am 339 305 90.0% 11 am – 12 pm 372 329 88.4% 12 pm – 1 pm 435 377 86.7% 1 pm – 2 pm 515 472 91.7% 2 pm – 3 pm 325 292 89.8% 3 pm – 4 pm 348 326 93.7% 4 pm – 5 pm 295 271 91.9% 5 pm – 6 pm 220 194 88.2% 6 pm – 7 pm 244 217 88.9%
Total 3,607 3,242 89.9%
Gender Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
Male 2,848 2,542 89.3% Female 750 691 92.1%
Unknown 9 9 100.0% Total 3,607 3,242 89.9%
Age Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
0 - 15 53 50 94.3% 16 - 29 725 636 87.7% 30 – 59 2,339 2,116 90.5%
60+ 490 440 89.8% Unknown 0 0 N/A
Total 3,607 3,242 89.9%
Race Actual Total No. of
Observations Actual Belted No. of
Observations % Safety Belt Use
White 3,315 2,981 89.9% Black 164 145 88.4% Other 122 112 91.8%
Unknown 6 4 66.7% Total 3,607 3,242 89.9%
26
Occupants of sport utility vehicles exhibited the highest safety belt use rate among vehicle types at 94.9
percent, followed closely by occupants of vans or minivans at 94.8 percent. Occupants of passenger cars
exhibited a use rate of 93.4 percent, while occupants of pick-up trucks exhibited the lowest use rate at 89.9
percent; consistent with historical trends. Considering days of the week, Wednesdays demonstrated the
lowest safety belt usage rate with 91.4 percent. Safety belt use rates were highest on Sundays with a rate
of 95.3 percent. The time period of 8:00 AM to 9:00 AM exhibited a lower usage rate than all other times
of the day (91.4 percent), while occupants were most likely to wear their safety belts between the hours of
1:00 PM to 2:00 PM (94.5 percent).
Female occupants had higher use rates than male occupants by 2.6 percent (94.9 percent use rate for
females vs. 92.3 percent use rate for males). The safety belt usage rate was highest among occupants
age 60 and above at 94.4 percent, and lowest for occupants between the ages of 16 to 29 (91.9 percent).
The safety belt use rate for occupants age 0 to 15 was found to be 93.7 percent while the use rate was
94.0 percent among occupants ages 30 to 59. Considering occupant races, the safety belt use rate was
found to be lowest among black occupants (91.0 percent) and highest for individuals of ‘other’ races (94.2
percent) which includes individuals of Asian descent and Pacific Islanders. White occupants were found to
have a safety belt use rate of 93.8 percent.
Table 15 summarizes occupant safety belt use rates by gender, age, and race. Vehicle occupants whose
gender could not be identified were excluded from this demographic comparison. Black males ages 16 to
29 exhibited a low belt use rate of 86.5%. However it should be noted that the sample size for this group
was relatively small. Similar to previous findings, white females and females of ‘other’ races of all ages
generally exhibited the highest safety belt use rates compared with other demographics. Additionally,
young male pick-up truck occupants exhibited the low safety belt use rates (89.3% for all male pickup truck
occupants, and 87.7% for all pickup truck occupants ages 16 to 29), consistent with past findings.
27
Table 15. Annual Belt Use by Demographic Characteristics
Demographic Data All Vehicles Safety Belt Use
Gender Age Race Actual Total No. of
Observations Actual Belted No. of Observations
% Safety Belt Use
Male
0 - 15
White 138 129 93.5% Black 26 24 92.3% Other 20 20 100.0%
Unknown 1 0 0.0% Total 185 173 93.5%
16 - 29
White 2,071 1,886 91.1% Black 386 334 86.5% Other 247 231 93.5%
Unknown 4 3 75.0% Total 2,708 2,454 90.6%
30 - 59
White 5,503 5,113 92.9% Black 661 608 92.0% Other 350 325 92.9%
Unknown 17 16 94.1% Total 6,531 6,062 92.8%
60+
White 1,676 1,557 92.9% Black 49 42 85.7% Other 20 20 100.0%
Unknown 0 0 N/A Total 1,745 1,619 92.8%
Unknown
White 3 3 100.0% Black 1 1 100.0% Other 0 0 N/A
Unknown 0 0 N/A Total 4 4 100.0%
TOTAL 11,173 10,312 92.3%
28
Table 15. Annual Belt Use by Demographic Characteristics (Continued)
Demographic Data All Vehicles Safety Belt Use
Gender Age Race Actual Total No. of
Observations Actual Belted No. of Observations
% Safety Belt Use
Female
0 - 15
White 114 108 94.7% Black 22 19 86.4% Other 12 12 100.0%
Unknown 0 0 N/A Total 148 139 93.9%
16 - 29
White 2,278 2,131 93.5% Black 374 336 89.8% Other 178 171 96.1%
Unknown 2 2 100.0% Total 2,832 2,640 93.2%
30 - 59
White 4,118 3,945 95.8% Black 658 617 93.8% Other 213 201 94.4%
Unknown 4 3 75.0% Total 4,993 4,766 95.5%
60+
White 1,295 1,249 96.4% Black 39 37 94.9% Other 20 19 95.0%
Unknown 0 0 N/A Total 1,354 1,305 96.4%
Unknown
White 2 2 100.0% Black 2 2 100.0% Other 0 0 N/A
Unknown 0 0 N/A Total 4 4 100.0%
TOTAL 9,331 8,854 94.9%
In comparison to 2017, the 2018 Annual survey revealed a slight decrease in safety belt usage from 94.1
percent to 93.4 percent; a change that is not statistically significant. In any case, continued public
awareness and enforcement efforts are warranted to increase safety belt use. The careful evaluation of
these media and enforcement efforts will allow for the identification of at-risk vehicle occupants and
geographic areas prone to low belt use rates. As shown in this and previous studies, young males and
pick-up truck drivers continue to exhibit lower safety belt use rates. Generally, belt use was also lower for
those counties in Stratum 1. These areas should be emphasized in subsequent program efforts.
29
6.2 Mobile Device Use Results and Conclusions
As a part of the 2018 annual observational survey of safety belt use, mobile device use was also recorded
for drivers only (passengers were not observed for mobile device use). A total of 1,154 drivers were
observed using a mobile device in some way and the overall weighted mobile device use rate was found
to be 7.1 percent. The weighted mobile device use rate (shown in Table 16) was calculated using the same
procedure as the weighted safety belt rate described in the “Data Analysis” section of the report. This rate
represents a 1.1 percent increase from the 6.0 percent mobile device use rate observed in Michigan in
2017. Nationally, the overall mobile device use rate by drivers was found to be 5.9 percent in 2016 [7] (the
most recent national data available), which included hand-held talking, hands-free talking (earpiece
observed), and typing, although hands-free devices with no earpiece observed were not included.
Michigan’s weighted mobile device use rate of 7.1 percent is slightly higher than the national average of
5.9 percent. Table 17 presents overall driver mobile device use, in addition to mobile device use by device
type and type of use.
Table 16. Annual Weighted Mobile Device Use Rate for Drivers
Use by Category Use Rate* Standard Error
Overall Mobile Device Use 7.1% ± 0.8% 0.4%
* Weighted Mobile Device Usage 95% Confidence Band
Table 17. Annual Unweighted Mobile Device Use Rates by Use Type
Use by Category Total # of Driver
Observations
Total # of Drivers Observed Using Mobile Device
Percent of Mobile Device Use by
Type (Drivers)
Talking – Hand-held Device 16,738 664 4.0%
Talking – Hands-free Device (Earpiece Observed)
16,738 68 0.4%
Talking – Hands-free Device (Earpiece Not Observed)
16,738 40 0.2%
Typing – Hand-held 16,738 382 2.3%
Overall Mobile Device Use 16,738 1154 6.9%
Table 18 summarizes mobile device use for drivers in terms of day of the week, time of the day, vehicle
type, gender, age and race. Females were found to be more likely to use a mobile device while driving than
males (8.2 percent and 6.0 percent, respectively). The mobile device use rate was found to be highest
between 4:00 pm and 5:00 pm at 7.9 percent, while the mobile device use rate was lowest between 5:00
pm and 7:00 am and 8:00 am and (5.4 percent). Mobile device use among drivers less than 30 years of
age was greatest at 10.2 percent, in comparison to 6.7 percent among those between ages 30 and 59 and
30
1.9 percent for drivers age 60 and above. Additionally, black drivers tended to exhibit higher mobile device
use rates while driving as compared to other demographics. Turning to days of the week, mobile device
use was highest on Wednesdays (8.3%), and lowest on Sundays (3.2%). Finally, mobile device use was
highest among drivers of vans/minivans (7.4%), and lowest among drivers of pickup trucks (6.1%).
Table 18. Annual Mobile Device Use Summary
Day of the Week
All Vehicles Mobile Device Use
Total No. of Driver
Observations
Total No. of Drivers
Observed Using Mobile Device
Percent of Mobile Device Use
(Drivers)
Sunday 1,851 67 3.6%
Monday 3,518 253 7.2%
Tuesday 2,959 217 7.3%
Wednesday 2,283 190 8.3%
Thursday 2,238 167 7.5%
Friday 1,749 128 7.3%
Saturday 2,140 132 6.2%
Total 16,738 1,154 6.9%
Time of the Day
All Vehicles Mobile Device Use
Total No. of Driver
Observations
Total No. of Drivers
Observed Using Mobile Device
Percent of Mobile Device Use
(Drivers)
7 am - 8 am 626 34 5.4%
8 am - 9 am 992 63 6.4%
9 am - 10 am 920 65 7.1%
10 am - 11 am 1,438 92 6.4%
11 am - 12 pm 1,848 108 5.8%
12 pm - 1 pm 2,148 165 7.7%
1 pm - 2 pm 2,021 135 6.7%
2 pm - 3 pm 1,592 115 7.2%
3 pm - 4 pm 1,667 115 6.9%
4 pm - 5 pm 1,334 106 7.9%
5 pm - 6 pm 1,138 88 7.7%
6 pm - 7 pm 1,014 68 6.7%
Total 16,738 1,154 6.9%
31
Table 18. Annual Mobile Device Use Summary (Continued)
Vehicle Type
All Vehicles Mobile Device Use
Total No. of Driver
Observations
Total No. of Drivers
Observed Using Mobile Device
Percent of Mobile Device Use
(Drivers)
Passenger Cars 6,248 424 6.8%
Sport Utility Vehicles 6,060 439 7.2%
Vans/ Minivans 1,508 112 7.4%
Pick-Up Trucks 2,922 179 6.1%
Total 16,738 1,154 6.9%
Gender
All Vehicles Mobile Device Use
Total No. of Driver
Observations
Total No. of Drivers
Observed Using Mobile Device
Percent of Mobile Device Use
(Drivers)
Male 9,786 585 6.0%
Female 6,917 565 8.2%
Unknown 35 4 11.4%
Total 16,738 1,154 6.9%
Age
All Vehicles Mobile Device Use
Total No. of Driver
Observations
Total No. of Drivers
Observed Using Mobile Device
Percent of Mobile Device Use
(Drivers)
16-29 4,400 447 10.2%
30-59 9,914 661 6.7%
60+ 2,415 45 1.9%
Unknown 9 1 11.1%
Total 16,738 1,154 6.9%
Race
All Vehicles Mobile Device Use
Total No. of Driver
Observations
Total No. of Drivers
Observed Using Mobile Device
Percent of Mobile Device Use
(Drivers)
White 14,046 873 6.2%
Black 1,854 220 11.9%
Other 809 56 6.9%
Unknown 29 5 17.2%
Total 16,738 1,154 6.9%
32
REFERENCES
1. Early Estimate of Motor Vehicle Traffic Fatalities for the First Half (Jan-Jun) of 2017. Rep. no. DOT
HS 812 453. Washington DC: National Highway Traffic Safety Administration, 2017.
2. NHTSA’s National Center for Statistics and Analysis, “Traffic Safety Facts - 2016 Data – Occupant
Protection in Passenger Vehicles”, U.S. Department of Transportation, NHTSA, DOT HS 812 494,
February 2018.
3. Blincoe, L. J., Miller, T. R., Zaloshnja, E., Lawrence, B. A. The Economic and Societal Impact of
Motor Vehicle Crashes, 2010 (Revised). Washington, DC: National Highway Traffic Safety
Administration, 2015.
4. Seat Belt Use in 2017 – Overall Results (Revised). (Traffic Safety Facts Research Note. Report
No. DOT HS 812 465). Washington, DC: National Highway Traffic Safety Administration, April
2018.
5. Seat Belt Use in 2017 - Use Rates in the States and Territories (Revised). Rep. no. DOT HS 812
546. Washington DC: National Highway Traffic Safety Administration, June 2018.
6. National Highway Traffic Safety Administration, An Example of a Compliant State Seat Belt Use
Survey Design, DOT HS 811 494, June 2011.
7. Driver Electronic Device Use in 2016. (Traffic Safety Facts Research Note. Report No. DOT HS
812 426). Washington, DC: National Highway Traffic Safety Administration, June 2017.
33
APPENDIX I Michigan Safety Belt Survey Cover Sheet and Data Collection Form
34
DIRECT OBSERVATION SURVEY COVER SHEET Date: _______ - _______ - 2018 Observer’s Name:__________________________ Site Identification: Site Location: _________________________________________________________ Site Number: City___________________________County____________________________Stratum_____ Alternate Site Information: Is this an alternate site? No Yes (Circle one) If yes, please provide a reason for using an alternate site from the reserve list: ____________________________________________________________________ Site Description: Observation direction: Northbound Southbound Eastbound Westbound
Number of lanes observed: ____________
Total number of lanes in this direction: ____________
Weather Conditions: Clear Light Fog Light Rain Site Start and End Time: Start time: ______________am/pm End time: _______________am/pm Sample Size
60 Minute Volume Count (for lanes being observed): ___________Vehicles
Number of Observations Recorded in 60 min: ___________Vehicles
35
OBSERVATION DATA COLLECTION SHEET
Note: E.P. = Ear Piece
36
APPENDIX II Resumes of Timothy J. Gates and Peter T. Savolainen
37
Dr. Timothy J. Gates Summary
Dr. Timothy J. Gates is the current Principal Investigator of the Direct Observation Survey of Safety Belt Use. Dr. Gates is an Associate Professor in the Michigan State University (MSU) Department of Civil and Environmental Engineering. He has more than eight years of experience with direct observation surveys of safety restraint use. This includes a diverse range of experiences in sample design and selection, field data collection methods, observer training, statistical systems development, and optimization techniques. He also has expertise in the areas of survey research methodology, data processing, and statistical quality control.
Education
Ph.D., Civil Engineering, University of Wisconsin, 2007 M.A., Civil Engineering, Michigan State University, 2000 B.S., Civil Engineering, Michigan State University, 2000
Professional Associations
American Society of Civil Engineers Institute of Transportation Engineers
Computer Skills
Operation Systems: Windows, iOs Software: LIMDEP, SAS, SPSS, SUDAAN, Microsoft PowerPoint, Excel and Word
Relevant Project Experience Wayne State University (2007 to Present)
Direct Observation Surveys of Seat Belt Use –PI or co-PI on OHSP-sponsored Michigan safety belt use survey from FY 2012 to present. Participated in proposal development, planning, survey implementation, data collection, quality control, data analysis, and report preparation. Led the resampling of Michigan’s 200 safety belt observation sites for use beginning with the 2018 survey. Direct Observation Surveys of Commercial Motor Vehicle Seat Belt Use – Co-PI on OHSP-sponsored Michigan seat belt use survey for commercial motor vehicle occupants during FY 2012 and 2015. Direct Observation Surveys of Child Restraint Device Use and Misuse (including Booster Seat Use) – PI or co-PI on OHSP-sponsored child restraint device use/misuse survey, including booster seats in FY 2009, 2011, 2013, 2015, and 2018. Direct Observation Surveys of Motorcycle Helmet Use – co-PI on OHSP-sponsored motorcycle helmet use survey in FY 2013 and 2017.
38
Dr. Peter T. Savolainen Summary
Dr. Peter T. Savolainen is an Associate Professor in the Iowa State University Department of Civil, Construction, and Environmental Engineering. Dr. Savolainen serves as the lead statistical advisor for this project. Prior to joining Iowa State University in 2014, he was an Associate Professor of Civil Engineering at Wayne State University. He has more than nine years of experience with direct observation surveys of safety restraint use. This includes a diverse range of experiences in sample design and selection, data weighting, imputation, variance estimation, statistical systems development, and optimization techniques. He also has expertise in the areas of survey research methodology, data processing, and statistical quality control. Dr. Savolainen also teaches graduate level courses on civil engineering research methods and applications, as well as statistics and econometric methods of data analysis. He is a proficient user of various statistical analysis software packages, including LIMDEP, SAS, SPSS, and SUDAAN.
Education
Ph.D., Civil Engineering, Purdue University, 2006 M.A., Civil Engineering, Purdue University, 2004 B.S., Civil Engineering, Michigan Technological University, 2002
Professional Associations
American Society of Civil Engineers American Statistical Association Institute of Transportation Engineers
Computer Skills
Operation Systems: Windows, iOs Software: LIMDEP, SAS, SPSS, SUDAAN, Microsoft PowerPoint, Excel and Word
Relevant Project Experience Wayne State University (2006 to Present)
Direct Observation Surveys of Seat Belt Use –PI or co-PI on OHSP-sponsored Michigan safety belt use survey from FY 2008 to 2010 and FY 2012 to present. Participated in proposal development, planning, survey implementation, data collection, quality control, data analysis, and report preparation. Led development of the federally-approved safety belt observational survey methodology for the state of Michigan in 2012. Direct Observation Surveys of Commercial Motor Vehicle Seat Belt Use – Co-PI on OHSP-sponsored Michigan seat belt use survey for commercial motor vehicle occupants during FY 2012. Direct Observation Surveys of Child Restraint Device Use and Misuse (including Booster Seat Use) – PI or co-PI on OHSP-sponsored child restraint device use/misuse survey, including booster seats in FY 2009, 2011, 2013, and 2015. Direct Observation Surveys of Motorcycle Helmet Use – co-PI on OHSP-sponsored motorcycle helmet use survey in FY 2013.
39
APPENDIX III
List of Annual Observation Locations by County, Stratum, and Road Classification Including Belt
Use Observation Data
40
41
42
43